Social Network User Subgroup Description Generation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Organizations face difficulties in identifying and describing user sub-groups within larger user groups on social networking systems, as existing methods lack a standardized way to convey meaningful information about sub-group composition and interests.
Innovation Solution
The solution involves clustering users into sub-groups using feature vectors and distance functions, ranking objects based on sub-group characteristics, and generating descriptions by combining the topics associated with highest-ranking pages, which are computed using conditional probabilities and intermediate variables like user interests.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If clustering algorithms use traditional feature vectors with high sparsity to represent users, then the representation captures detailed user-object affinities, but the clustering ability deteriorates due to the sparsity making it difficult to meaningfully cluster users
Solution Approach 1:
The patent transforms the feature representation from a high-dimensional sparse vector of object affinities to a lower-dimensional dense vector of inferred interests. This parameter transformation changes the nature of the data from sparse to dense, enabling effective clustering while preserving meaningful user characteristics. The transformation is achieved through matrix factorization or probabilistic topic models that map object affinities to latent interest dimensions.
2Loss of information
If sub-groups are identified without a standardized description method, then the sub-group composition can be analyzed, but the ability to convey meaningful information about sub-group characteristics deteriorates
Solution Approach 1:
The patent extracts key characteristic objects from each sub-group based on their ranking scores and uses these extracted objects to form human-readable descriptions. Instead of presenting all sub-group members or all their attributes, the system extracts the most representative objects (e.g., favorite pages, frequent interactions) and combines their topics to create concise sub-group descriptions that convey essential information about sub-group composition and interests.
3Device complexity
If traditional clustering methods are used on sparse user data, then the computational process is simpler, but the reliability of sub-group identification deteriorates due to inability to capture meaningful user similarities
Solution Approach 1:
The patent performs preliminary transformation of user data from sparse object affinity vectors to dense interest vectors before applying clustering algorithms. This preliminary action converts the data into a format that is more suitable for clustering, ensuring that meaningful user similarities are captured. By preprocessing the data to infer latent interests from observed affinities, the system ensures that subsequent clustering operations are performed on reliable, dense representations that accurately reflect user characteristics.
Data Source
AI summary
A system forms sub-groups from a given user group of a social networking system and form descriptions of the sub-groups that provide an intuitive understanding of sub-group composition, such as likings of the sub-groups. In one embodiment, a given user group of a social networking system is clustered into a plurality of sub-groups, and representative characteristics—such as the characteristics of a composite or actual member of the sub-group—are determined for each sub-group. In order to form sub-group descriptions, a set of objects, such as pages of the social networking system, is ranked with respect to the representative characteristics of the sub-group. The highest-ranking objects for a sub-group are then used to form the description of that sub-group. For example, the topics associated with each of the highest-ranking pages can be combined into the sub-group description.